This year, I’ve applied to U.S. colleges, which are notorious for their ultra-competitive admissions, and that is doubly true for international students. In the wake of the long waiting time after hitting the submit button on my application, constant doubts have crept into my mind. Are my extracurricular activities competitive enough? Are my essays unique and authentic? Are my awards truly selective?

These doubts are natural, and in the past, we could do nothing but wait patiently for the admission results to come out, and only then would our doubts be put to rest.

But now, I found myself going towards a sort of omniscient entity that could help alleviate such doubts. I booted up ChatGPT and shared with it all my academic statistics and all my essays and activities, and asked it the questions that troubled me. And like the subservient and sycophantic subordinate it is, it vehemently defended my application. 

That gave me confidence. It felt liberating for the AI to chase away all my fears and doubts. But the thing with doubts is that they keep on coming. Either they can be buried or embraced; they can’t be chased away. So, when the doubts would inevitably come back, I would again turn to Chat GPT for some over-the-top—bordering on deceitful—compliments about myself and my application.

This created a cycle of constant external validation for me, and I would find myself going to ChatGPT several times a day just to hear that my essays were thought-provoking, my activities were unique and interesting, and I was a smart and capable individual.

This reversed the dynamics. Now I was the subservient one in this interaction, craving for validation from AI. Realizing this toxic cycle, I instantly decided to go cold turkey. It was difficult in the beginning to keep myself from going back to hear the AI’s soothing affirmations. That just confirmed my fears: I was addicted. Soon, the spell was broken, thankfully, because I had noticed the patterns soon enough. 

Of course, we all know that AI is built to tell us what we want to hear. If you ask, ‘Is my writing good?’, it will reply with positive affirmations. While it might point out revisions and give actionable feedback, it will preface all of it by saying something like: “Your writing is already good; these are just some stylistic choices which might elevate it even beyond.”

With newfound confidence, you start posting on Reddit communities, not to get human feedback, but to get human compliments. The next thing you know, your piece is dubbed as the ‘worst thing they’ve read in their life’. This blunt—and even rude—feedback isn’t because you’re a bad writer, its just that when your first draft, which is meant to be terrible or at least doesn’t reflect your true ability as a writer, is praised to be the finest work of literature known to man, of course this false sense of confidence will begin to take root.

But the worst course of action is becoming so confident as to not even post it on Reddit and get your write-up torn to shreds. In that case, at least you're given some realistic feedback. But if you're so confident that you don't even think about cross-checking it, you might end up drowning in complacency because you believe you’ve already reached the pinnacle of literature and the end of your potential.

Therefore,

In this way, AI creates:

 1) A cycle of craving attention and affirmation, 

2) A false sense of confidence by giving us feedback which is, at the least, overblown and at the most, deceitful. 

These problems are distinct yet intertwined, and both of these hamper our long-term independent growth and self-improvement in any domain, whether it be analytical or creative. As such, young people need to be very aware of this because AI’s affirmation is one of the most potent addictions of the contemporary age.